In this paper we show how the action semantics framework can be used to describe a particular implementation of fault-tolerant systems. We also de ne a notion of simulation which can be the basis for relating a fault-tolerant implementation to an abstract (non-faulty, and non-fault tolerant) specification. The aim of the paper is to illustrate that good software engineering techniques can be applied to semantic descriptions. Issues such as modularity, extensibility of the semantic descriptions is illustrated.